Search

Magnetic resonance imaging (MRI) provides excellent soft-tissue contrast for clinical diagnoses and research which underpin many recent breakthroughs in medicine and biology. The post-processing of reconstructed MR images is often automated for incorporation into MRI scanners by the manufacturers and increasingly plays a critical role in the final image quality for clinical reporting and interpretation. For image enhancement and correction, the post-processing steps include noise reduction, image artefact correction, and image resolution improvements. With the recent success of deep learning in many research fields, there is great potential to apply deep learning for MR image enhancement, and recent publications have demonstrated promising results. Motivated by the rapidly growing literature in this area, in this review paper, we provide a comprehensive overview of deep learning-based methods for post-processing MR images to enhance image quality and correct image artefacts. We aim to provide researchers in MRI or other research fields, including computer vision and image processing, a literature survey of deep learning approaches for MR image enhancement. We discuss the current limitations of the application of artificial intelligence in MRI and highlight possible directions for future developments. In the era of deep learning, we highlight the importance of a critical appraisal of the explanatory information provided and the generalizability of deep learning algorithms in medical imaging.

Estimation of accurate attenuation maps for whole-body positron emission tomography (PET) imaging in simultaneous PET-MRI systems is a challenging problem as it affects the quantitative nature of the modality. In this study, we aimed to improve the accuracy of estimated attenuation maps from MRI Dixon contrast images by training an augmented generative adversarial network (GANs) in a supervised manner. We augmented the GANs by perturbing the non-linear deformation field during image registration between MRI and the ground truth CT images. We acquired the CT and the corresponding PET-MR images for a cohort of 28 prostate cancer patients. Data from 18 patients (2160 slices and later augmented to 270,000 slices) was used for training the GANs and others for validation. We calculated the error in bone and soft tissue regions for the AC μ-maps and the reconstructed PET images. For quantitative analysis, we use the average relative absolute errors and validate the proposed technique on 10 patients. The DL-based MR methods generated the pseudo-CT AC μ-maps with an accuracy of 4.5% more than standard MR-based techniques. Particularly, the proposed method demonstrates improved accuracy in the pelvic regions without affecting the uptake values. The lowest error of the AC μ-map in the pelvic region was 1.9% for μ-mapGAN + aug compared with 6.4% for μ-mapdixon, 5.9% for μ-mapdixon + bone, 2.1% for μ-mapU-Net and 2.0% for μ-mapU-Net + aug. For the reconstructed PET images, the lowest error was 2.2% for PETGAN + aug compared with 10.3% for PETdixon, 8.7% for PETdixon + bone, 2.6% for PETU-Net and 2.4% for PETU-Net + aug.. The proposed technique to augment the training datasets for training of the GAN results in improved accuracy of the estimated μ-map and consequently the PET quantification compared to the state of the art.

The suppression of motion artefacts from MR images is a challenging task. The purpose of this paper was to develop a standalone novel technique to suppress motion artefacts in MR images using a data-driven deep learning approach. A simulation framework was developed to generate motion-corrupted images from motion-free images using randomly generated motion profiles. An Inception-ResNet deep learning network architecture was used as the encoder and was augmented with a stack of convolution and upsampling layers to form an encoder-decoder network. The network was trained on simulated motion-corrupted images to identify and suppress those artefacts attributable to motion. The network was validated on unseen simulated datasets and real-world experimental motion-corrupted in vivo brain datasets. The trained network was able to suppress the motion artefacts in the reconstructed images, and the mean structural similarity (SSIM) increased from 0.9058 to 0.9338. The network was also able to suppress the motion artefacts from the real-world experimental dataset, and the mean SSIM increased from 0.8671 to 0.9145. The motion correction of the experimental datasets demonstrated the effectiveness of the motion simulation generation process. The proposed method successfully removed motion artefacts and outperformed an iterative entropy minimization method in terms of the SSIM index and normalized root mean squared error, which were 5-10% better for the proposed method. In conclusion, a novel, data-driven motion correction technique has been developed that can suppress motion artefacts from motion-corrupted MR images. The proposed technique is a standalone, post-processing method that does not interfere with data acquisition or reconstruction parameters, thus making it suitable for routine clinical practice.

Post-transcriptional non-template additions of nucleotides to 3′-ends of RNAs play important roles in the stability and function of RNA molecules. Although tRNA nucleotidyltransferase (CCA-adding enzyme) is known to add CCA trinucleotides to 3′-ends of tRNAs, whether other RNA species can be endogenous substrates of CCA-adding enzyme has not been widely explored yet. Herein, we used YAMAT-seq to identify non-tRNA substrates of CCA-adding enzyme. YAMAT-seq captures RNA species that form secondary structures with 4-nt protruding 3′-ends of the sequence 5′-NCCA-3′, which is the hallmark structure of RNAs that are generated by CCA-adding enzyme. By executing YAMAT-seq for human breast cancer cells and mining the sequence data, we identified novel candidate substrates of CCA-adding enzyme. These included fourteen ‘CCA-RNAs’ that only contain CCA as non-genomic sequences, and eleven ‘NCCA-RNAs’ that contain CCA and other nucleotides as non-genomic sequences. All newly-identified (N)CCA-RNAs were derived from the mitochondrial genome and were localized in mitochondria. Knockdown of CCA-adding enzyme severely reduced the expression levels of (N)CCA-RNAs, suggesting that the CCA-adding enzyme-catalyzed CCA additions stabilize the expression of (N)CCA-RNAs. Furthermore, expression levels of (N)CCA-RNAs were severely reduced by various cellular treatments, including UV irradiation, amino acid starvation, inhibition of mitochondrial respiratory complexes, and inhibition of the cell cycle. These results revealed a novel CCA-mediated regulatory pathway for the expression of mitochondrial non-coding RNAs.

To evaluate the clinical utility of the application of a deep learning motion correction technique on 3D MPRAGE magnetic resonance images acquired in routine clinical practice.An encoder-decoder deep learning network inspired by InceptionResnet was trained on public datasets. The clinical utility of the trained network was evaluated retrospectively on 27 3D MPRAGE T1 weighted motion degraded MR images identified by radiologists during reporting. The assessment of image quality was performed by one board-certified radiologist and one senior radiology trainee for nine neuroanatomical regions of the brain using a five-point visual grading scale.The deep learning motion correction technique resulted in reduced ghosting, ringing and blurring for all the brain regions investigated. The larger regions of interest such as ventricles improved the least (1.81 to 1.16, p-value:0.0001) while the smaller but complex regions such as the hippocampus improved most (3.0 to 1.67, p-value:0.0001). The Wilcox rank tests of image quality differences for the nine neuroanatomical regions were all statistically significant (p 0.001). Overall, 60 % of the neuroanatomical regions were improved, 39 % were unchanged and 1 % were degraded. Out of the unchanged cases, 28 % were already scored at the highest image quality before motion correction. It was found that approximately 13 % of repeated scans could be avoided using the DL motion correction approach.The deep learning motion correction technique improved the overall visual perception of the 3D T1 weighted MPRAGE brain images. This would improve the clinical utility of otherwise motion degraded images and allow visualisation of normal anatomy and even subtle pathology.

Estimation of accurate attenuation maps for whole-body positron emission tomography (PET) imaging in simultaneous PET-MRI systems is a challenging problem as it affects the quantitative nature of the modality. In this study, we aimed to improve the accuracy of estimated attenuation maps from MRI Dixon contrast images by training an augmented generative adversarial network (GANs) in a supervised manner. We augmented the GANs by perturbing the non-linear deformation field during image registration between MRI and the ground truth CT images.We acquired the CT and the corresponding PET-MR images for a cohort of 28 prostate cancer patients. Data from 18 patients (2160 slices and later augmented to 270,000 slices) was used for training the GANs and others for validation. We calculated the error in bone and soft tissue regions for the AC μ-maps and the reconstructed PET images.For quantitative analysis, we use the average relative absolute errors and validate the proposed technique on 10 patients. The DL-based MR methods generated the pseudo-CT AC μ-maps with an accuracy of 4.5% more than standard MR-based techniques. Particularly, the proposed method demonstrates improved accuracy in the pelvic regions without affecting the uptake values. The lowest error of the AC μ-map in the pelvic region was 1.9% for μ-map

An implementation of Non-Fourier chirp-encoding in 3D Gradient Recalled Echo (GRE), susceptibility-weighted imaging (SWI) and Magnetization Prepared Rapid Gradient Echo (MPRAGE) sequences is presented with compressive sensing reconstruction. 3D GRE and MPRAGE sequences were designed, in which the phase encoding (PE) direction was encoded with spatially selective chirp encoding Radio Frequency (RF) pulses, while the slice and the readout directions were Fourier encoded using gradients. During each excitation along the PE direction, a different spatially-selective RF excitation pulse was used to encode the PE direction with a complete set of unitary chirp encoding basis. Multichannel compressive sensing reconstruction on the undersampled in vivo data demonstrated that images reconstructed from chirp encoded data were able to preserve the spatial resolution better than the Fourier encoding. The mean Structural Similarity (SSIM) across five subjects at the acceleration factor of 6, for chirp encoded MPRAGE was 0.934 compared to 0.912 for Fourier encoded MPRAGE. The implementation of prospective undersampling demonstrated the feasibility of using chirp encoding in clinical practice for accelerated imaging. The minimum intensity projection of the compressive sensing (CS) reconstructed susceptibility weighted images revealed that chirp encoding is able to delineate small vessels better than the Fourier encoding with the SSIM of 0.960 for chirp encoding compared to the SSIM of 0.949 for the Fourier encoding. Improved performance of chirp encoding for CS reconstruction and SWI, along with the feasibility of implementation makes them a practical candidate for clinical MRI scans.

We propose an ensemble of 2D convolutional neural networks to predict the 3D brain tumor segmentation mask using the multi-contrast brain images. A pretrained Resnet50 and Nasnet-mobile architecture were used as an encoder, which was appended with a decoder network to create an encoder-decoder neural network architecture. The encoder-decoder network was trained end to end using T1, T1 contrast-enhanced, T2 and T2-Flair images to classify each pixel in the 2D input image to either no tumor, necrosis/non-enhancing tumor (NCR/NET), enhancing tumor (ET) or edema (ED). Separate Resent50 and Nasnet-mobile architectures were trained for axial, sagittal and coronal slices. Predictions from 5 inferences including Resnet at all three orientations and Nasnet-mobile at two orientations were averaged to predict the final probabilities and subsequently the tumor mask. The mean dice scores calculated from 166 were 0.8865, 0.7372 and 0.7743 for whole tumor, tumor core and enhancing tumor respectively.

A deep learning (DL) method for accelerated magnetic resonance (MR) imaging is presented that incorporates domain knowledge of parallel MR imaging to augment the DL networks for accurate and stable image reconstruction. The proposed DL method employs a novel loss function consisting of a combination of mean absolute error, structural similarity, and sobel edge loss. The DL model takes both original measurements and images reconstructed by the parallel imaging method as inputs to the network. The accuracy of the proposed method was evaluated using two anatomical regions and six MRI contrasts and was compared with state-of-the-art parallel imaging and deep learning methods. The proposed method significantly outperformed the other methods for all the six different contrasts in terms of structural similarity, peak signal to noise ratio, and normalized mean squared error. The out-of-sample performance of the proposed method was assessed for a truly “unseen” case in a volunteer scan. The method produced images without any artificial features, often occurring in the DL image reconstruction methods. A stability analysis was performed by adding perturbations to the input, which demonstrated that the proposed method is robust and stable with respect to small structural changes, and different undersampling ratios. Comprehensive validation on large datasets demonstrated that incorporation of domain knowledge sufficiently regularizes the DL based image reconstruction and produces accurate and stable image enhancement.

In the pathogenesis of asthma, inflammatory mediators cause structural and functional changes in resident airway cells including airway smooth muscle (ASM) cells. Here, we report that intranasal treatment of mice with house dust mite (HDM) highly upregulated the expression of tRNA half molecules in asthmatic lungs. The 5′-tRNA haves contain a 2′,3′-cyclic phosphate (cP) and thus belong to cP-containing RNAs (cP-RNAs). Capturing the whole cP-RNA transcriptome using cP-RNA-seq revealed the global upregulation of not only tRNA halves but also the cP-RNAs derived from mRNAs and rRNAs. Our investigation of the biological significance of the major upregulated 5′-tRNA half species in human primary ASM cells revealed that the 5′-tRNA halves repress autophosphorylation at Tyr416 of Src protein kinase, a key regulator of the cellular focal adhesion pathway, leading to reduction of ASM cellular focal adhesion. These results assign a novel role as a protein modification modulator to tRNA half molecules and unveil a tRNA-engaged pathway in the molecular pathogenesis of asthma.

A deep learning approach to glioma segmentation is presented. An encoder and decoder pair deep learning network is designed which takes T1, T2, T1-CE (contrast enhanced) and T2-Flair (fluid attenuation inversion recovery) images as input and outputs the segmented labels. The encoder is a 49 layer deep residual learning architecture that encodes the \(240\,\times \,240\,\times \,4\) input images into \(8\,\times \,8\,\times \,2048\) feature maps. The decoder network takes these feature maps and extract the segmented labels. The decoder network is fully convolutional network consisting of convolutional and upsampling layers. Additionally, the input images are downsampled using bilinear interpolation and are inserted into the decoder network through concatenation. This concatenation step provides spatial information of the tumor to the decoder, which was lost due to pooling/downlsampling during encoding. The network is trained on the BRATS-17 training dataset and validated on the validation dataset. The dice score, sensitivity and specificity of the segmented whole tumor, core tumor and enhancing tumor is computed on validation dataset. The mean dice score for whole tumor, core tumor and enhancing tumor for validation dataset were 0.824, 0.627 and 0.575, respectively.

Contents Chromosomal fragmentations or damage in sperm DNA has considerable value in determination of semen quality. However, rapid and/or simple method to assess sperm DNA integrity in buffalo has apparently not been reported. In the present study, SCD was used for the first time in buffalo bulls for assessment of sperm DNA fragmentation. A modified SCD protocol, under bright field microscope was developed and validated by comparison with other routine tests which can be used for processing of samples. The DNA fragmentation index (DFI) from SCD was correlated with semen quality parameters viz. viability (r = −0.68, p

The incoherence between measurement and sparsifying transform matrices and the restricted isometry property (RIP) of measurement matrix are two of the key factors in determining the performance of compressive sensing (CS). In CS-MRI, the randomly under-sampled Fourier matrix is used as the measurement matrix and the wavelet transform is usually used as sparsifying transform matrix. However, the incoherence between the randomly under-sampled Fourier matrix and the wavelet matrix is not optimal, which can deteriorate the performance of CS-MRI. Using the mathematical result that noiselets are maximally incoherent with wavelets, this paper introduces the noiselet unitary bases as the measurement matrix to improve the incoherence and RIP in CS-MRI, and presents a method to design the pulse sequence for the noiselet encoding. This novel encoding scheme is combined with the multichannel compressive sensing (MCS) framework to take the advantage of multichannel data acquisition used in MRI scanners. An empirical RIP analysis is presented to compare the multichannel noiselet and multichannel Fourier measurement matrices in MCS. Simulations are presented in the MCS framework to compare the performance of noiselet encoding reconstructions and Fourier encoding reconstructions at different acceleration factors. The comparisons indicate that multichannel noiselet measurement matrix has better RIP than that of its Fourier counterpart, and that noiselet encoded MCS-MRI outperforms Fourier encoded MCS-MRI in preserving image resolution and can achieve higher acceleration factors. To demonstrate the feasibility of the proposed noiselet encoding scheme, two pulse sequences with tailored spatially selective RF excitation pulses was designed and implemented on a 3T scanner to acquire the data in the noiselet domain from a phantom and a human brain.

BACKGROUND: Salmonella are able to modulate host cell functions facilitating both uptake and resistance to cellular host defence mechanisms. While interactions between bacterial modulators and cellular proteins have been the main focus of Salmonella research, relatively little is known about mammalian gene regulation in response to Salmonella infection. A major class of mammalian gene modulators consists of microRNAs. For our study we examined interactions of microRNAs and regulated mRNAs in mammalian intestinal Salmonella infections using a piglet model. METHODOLOGY/PRINCIPAL FINDINGS: After performing microRNA as well as mRNA specific microarray analysis of ileal samples from Salmonella infected as well as control piglets, we integrated expression analysis with target prediction identifying microRNAs that mainly regulate focal adhesion as well as actin cytoskeleton pathways. Particular attention was given to miR-29a, which was involved in most interactions including Caveolin 2. RT-qPCR experiments verified up-regulation of miR-29a after infection while its predicted target Caveolin 2 was significantly down-regulated as examined by transcript and protein detection. Reporter gene assays as well as RNAi experiments confirmed Caveolin 2 to be a miR-29a target. Knock-down of Caveolin 2 in intestinal epithelial cells resulted in retarded proliferation as well as increased bacterial uptake. In addition, our experiments showed that Caveolin 2 regulates the activation of the small Rho GTPase CDC42 but apparently not RAC1 in human intestinal cells. CONCLUSIONS/SIGNIFICANCE: Our study outlines for the first time important regulation pathways in intestinal Salmonella infection pointing out that focal adhesion and organisation of actin cytoskeleton are regulated by microRNAs. Functional relevance is shown by miR-29a mediated Caveolin 2 regulation, modulating the activation state of CDC42. Further analysis of examined interactions may support the discovery of novel strategies impairing the uptake of intracellular pathogens.

The present study has examined the effect of different concentrations (1 μg/ml, 10 μg/ml and 100 μg/ml) of titanium oxide (TiO2) nanoparticles (NPs) (2 NP albeit at lower concentrations. In DNA fragmentation assay, there was dose-dependent increase in the DNA fragmentation ( r = 0.96). Ultrathin cross-sections revealed TiO2 NPs inside head and plasma membrane of the buffalo spermatozoa as assessed by TEM. These studies suggest that TiO2 NPs may have cytotoxic effect on buffalo spermatozoa by affecting sperm functionality and causing high amount of DNA fragmentations.